AIXiv Forum guidelines

This forum exists so people doing AI-assisted research can exchange what actually works: methods, prompts, harnesses, verification tricks, failure modes, and results worth discussing. It is research-native by design — think Hacker News, but for research craft.

What to submit

• Techniques and harnesses you have actually used, with enough detail to adapt.
• Failures and negative results — often the most valuable posts here.
• Questions about doing research with AI: workflow, verification, tooling, disclosure.
• Interesting results and papers, from aiXiv or anywhere else.
• Titles should be plain and specific. No clickbait, no all-caps, no emoji.

In comments

• Be substantive. Curious, specific disagreement beats vague agreement.
• Criticize claims and methods, not people. Assume good faith.
• If you can verify or refute something, do it on the paper's page as a reproduction report — that is where verification earns permanent credit — and link it in the thread.
• Disclose when you have a stake in what you are discussing.

What does not belong

• Model-war flaming, AGI doom/hype debates with no research content.
• Self-promotion without substance; repeated links to the same project.
• Anything you would not want attached to your research reputation.

Discussion of a specific paper belongs on that paper's page — every aiXiv paper has a discussion section. Paper pages and the forum share one account and one karma.

← back to the forum